Essential roles for deubiquitination in Leishmania life cycle progression

The parasitic protozoan Leishmania requires proteasomal, autophagic and lysosomal proteolytic pathways to enact the extensive cellular remodelling that occurs during its life cycle. The proteasome is essential for parasite proliferation, yet little is known about the requirement for ubiquitination/deubiquitination processes in growth and differentiation. Activity-based protein profiling of L. mexicana C12, C19 and C65 deubiquitinating cysteine peptidases (DUBs) revealed DUB activity remains relatively constant during differentiation of procyclic promastigote to amastigote. However, when life cycle phenotyping (bar-seq) was performed on a pool including 15 barcoded DUB null mutants created in promastigotes using CRISPR-Cas9, significant loss of fitness was observed during differentiation and intracellular infection. DUBs 4, 7, and 13 are required for successful transformation from metacyclic promastigote to amastigote and DUBs 3, 5, 6, 8, 10, 11 and 14 are required for normal amastigote proliferation in mice. DUBs 1, 2, 12 and 16 are essential for promastigote viability and the essential role of DUB2 in establishing infection was demonstrated using DiCre inducible gene deletion in vitro and in vivo. DUB2 is found in the nucleus and interacts with nuclear proteins associated with transcription/chromatin dynamics, mRNA splicing and mRNA capping. DUB2 has broad linkage specificity, cleaving all the di-ubiquitin chains except for Lys27 and Met1. Our study demonstrates the crucial role that DUBs play in differentiation and intracellular survival of Leishmania and that amastigotes are exquisitely sensitive to disruption of ubiquitination homeostasis.


Introduction
Leishmaniasis is a vector-borne neglected tropical disease that causes serious public health problems in 98 tropical and sub-tropical countries and is responsible for 20,000-40,000 deaths annually [1]. The disease spectrum varies from asymptomatic infection, through a disfiguring and debilitating mucocutaneous form to a visceral form that can be fatal. The causative agent of this disease is the obligate intracellular parasite of the genus Leishmania. The developmental process of Leishmania is closely linked with the ability of this parasite to survive and replicate in distinct environmental niches. The medical importance of one of these environmental niches is evident in the phagolysosomes of the infected host macrophages, where the Leishmania parasites differentiate from the motile promastigote to the non-motile amastigote and subsequently proliferate and establish an infection. The morphological and transcriptional changes that occur after differentiation of Leishmania promastigotes to amastigotes are well established [2], but little is known about the regulatory and operational processes that underpin this transition. Reversible post-translational modification (PTM) of proteins with chemical groups, such as phosphorylation, or other proteins, such as ubiquitination, are likely to be pivotal for successful life cycle progression [3][4][5] and intracellular parasitism. In addition, irreversible PTM such as proteolytic cleavage is important for the cellular remodeling that occurs during the life cycle of Leishmania and the commitment to differentiation [6][7][8], as well as in amastigote survival [9,10].
Ubiquitination is an evolutionarily conserved post-translational modification involving a series of highly regulated enzymatic reactions resulting in attachment of ubiquitin to proteins in the cell [11]. Ubiquitination regulates vital cellular mechanisms including protein degradation through the ubiquitin-proteasome system (UPS), autophagy, DNA repair and protein trafficking. For the covalent attachment of ubiquitin to target protein or other ubiquitin, three types of enzymes are required to act sequentially: E1 ubiquitin-activating enzyme, E2 ubiquitin-conjugating enzyme and an E3 ubiquitin ligase [11]. This system is reversible and deubiquitinase enzymes (DUBs) are responsible for the removal of ubiquitin from the target protein, providing an additional level of regulation of the ubiquitination pathway [12]. The number of DUBs varies from organism to organism. For example, around 100 DUBs have been discovered in humans but only 20 DUBs have been found in Saccharomyces cerevisiae [13]. In general, DUBs can be divided into seven distinct structural families [14]: ubiquitinspecific proteases (USPs, family C19), C-terminal hydrolyses (UCHs, family C12), ovarian tumor proteases (OTUs, family C65), JAB1/MPN/MOV34 metalloenzymes (JAMM/MPN+, family M67), Josephins, family C86, MINDY, family C115 [15] and the newly identified ZUFSP (zinc finger with UFM1-specific peptidase domain protein), family C78 [13,15]. Here

Leishmania DUBs required for life cycle transition
The fluorescent activity-based probe (ABP) Cy5UbPRG, previously developed for its ability to react with active DUBs [22], identified active DUBs in the lysate of L. mexicana promastigotes. In-gel fluorescence (Cy5) showed the presence of several DUBs in the range of 30-160 kDa (Fig 2A), the size expected taking into account the covalent attachment of the 9 kDa Cy5UbPRG probe (S1 Table). The activity of DUBs detected with the ABP was similar in promastigotes, axenic amastigotes and lesion-derived amastigotes and during differentiation from amastigotes to promastigotes (Fig 2B). UbPRG was then attached to Sepharose beads and used to affinity purify proteins from an L. mexicana promastigote cell lysate, and mass spectrometry used to confirm DUB identity. Six DUBs (DUB2, DUB15, DUB16, DUB17, DUB18, and DUB19) were successfully identified only in the UbPRG sample and not in the control sample, indicating that the warhead can successfully react with cysteine peptidase DUBs (Fig 2C and  S2A Table). Twenty other proteins specifically identified in the UbPRG sample are likely to be proteins that are DUB substrates (such as polyubiquitin, LmxM.09.0891), form complexes with DUBs, or react directly with the PRG warhead.
To assess whether the 20 DUBs were dispensable for growth of L. mexicana promastigotes each gene was targeted for deletion using CRISPR-Cas9 [23] with two barcoded antibiotic resistance cassettes to attempt to replace both allelic copies of each DUB (Fig 2C). In the first
To identify the DUB activities detected in promastigotes (Fig 2A), cell lysates for the DUB null mutants were probed with the Cy5UbPRG ABP ( Fig 2E, S3 Fig), with the expectation that the absence of a band of activity in comparison to the L. mexicana Cas9 T7 would represent the DUB that had been deleted. This allowed us to identify 5 active DUBs close to their expected sizes; DUB6 (133 kDa), DUB9 (155 kDa), DUB15 (34 kDa), DUB17 (30 kDa) and DUB18 (102 kDa). The fact that DUB15, DUB16, DUB17, DUB18, and DUB19 had been affinity purified with the UbPRG probe shows that these DUBs are active in the promastigote stage. A similar-sized DUB activity present in L. mexicana Cas9 T7 was absent in promastigotes of both Δdub13 and Δdub19, but we were unable to identify the DUB. Of note, PAN2, the human orthologue of DUB13, does not appear to bind the Cy5UbPRG probe [24], as it lacks key catalytic residues including the active Cys. This may also be true for L. mexicana DUB13.
To investigate further, promastigote cell lines were differentiated to amastigotes and probed with the ABP. This allowed DUB8 (172kDa) and DUB19 (148kDa) to be identified by comparing wild type and Δdub8 and Δdub19 (S3D Fig). Several null mutant DUB cell lines did not show any depletion of activity with the Cy5UbPRG profiling in either promastigotes or amastigotes, suggesting that those DUBs are not expressed, not active or are below detection limits of the ABP in the promastigote stage under the environmental conditions tested. Alternatively, the DUBs may not bind the ABP.
It was not possible to isolate null mutant clones for DUB1, DUB2, DUB12 and DUB16, possibly because those DUBs are essential in promastigotes ( Fig 2C). In three instances (DUB1, DUB2 and DUB12), correct integration of the drug resistance cassettes was observed, but the parasites retained a wild type copy of the gene, indicating that the genes are essential (S2 Fig). For DUB16, no populations could be recovered after the transfections. To provide further evidence that the genes are essential and to discount the possibility that the lack of gene deletion was a technical failure, facilitated null mutants were generated for the four DUB genes. Initially, the parental Cas9 T7 cell line was transfected with a plasmid expressing the DUB, thereafter the two alleles in the genomic locus were deleted successfully using CRISPR, two independent populations isolated, and the expected gene editing confirmed by diagnostic PCR (Fig 2D and S2C Fig). The Cy5UbPRG ABP was used to identify the over-expressed and active DUBs in the L. mexicana Cas9 T7 [DUB] cell lines (Fig 2F), where an increase in specific activities were detected in the over-expression lines compared to the parental line ( Fig 2F,

PLOS PATHOGENS
Leishmania deubiquitinases 45 and 50 kDa) are smaller than expected and seem likely to be proteolytically processed in the cell (Fig 2F).

Bar-seq reveals the importance of DUBs in the L. mexicana life cycle
To assess the ability of the DUB null mutants to transition through the parasite life cycle, barseq analysis, a method that involves the parallel phenotyping of pools of mutants using highthroughput DNA sequencing, was carried out [25]. Our 58 barcoded null mutants representing 16 DUBs, 13 other peptidases, 25 other ubiquitination system genes and 4 protein kinases (S4 Table) were prepared in equal proportions in 6 replicate promastigote (PRO) pools. The pools were grown for 7 days (d) before samples were collected and either induced to form axenic amastigotes (AXA) or used to purify metacyclic stage promastigotes (META). Purified metacyclic promastigotes were used to infect macrophages (inMAC) in vitro or mice via footpad (FP) injection. At various time points, DNA samples were prepared from the pool for amplification of the barcodes by PCR and quantitative analysis via next generation sequencing ( Fig 3A, S4 Table).
In order to analyse fitness, defined as "the reproductive capacity of a given population in a given environment" the number of reads for each barcode was first divided by the total number of reads for expected barcodes at each experimental stage. Next, data were averaged across the 6 replicates and the barcodes matched back to their respective null mutant lines. Changes in the average proportional representation of these lines were then used to infer changes in fitness. Importantly, because of the potential for the cumulative loss of null mutant lines during the experiment, statistical tests to investigate loss of fitness were only applied between adjacent time points. This ensured that the experimental period linked with a particular viability defect could be identified. Additionally, because of the higher likelihood that null mutant lines would exhibit decreased, rather than increased, fitness during the experiment (leading to an overall increase in the proportional representation of remaining null mutant lines), increases in fitness could not be confidently reported. Consequently, our analysis focuses only on deficiencies in life cycle progression.
Barcodes were detected for all null mutant lines at time point zero, however, very few reads were found for Δdub18 (most likely due to experimental error), leading this null mutant line to be excluded from the analysis. Decreased fitness was observed between adjacent promastigote stages (PRO 0 h-PRO 168 h) for a number of null mutant lines, with Δdub8, Δdub9 and Δdub13 in particular showing reduced fitness at two promastigote time points (unpaired ttests with Holm-Šídák correction for multiple comparisons, p<0.05). None of the mutants had significantly decreased fitness between the 168 h and metacyclic promastigote (META) samples, however, strong loss of fitness was common upon entering the amastigote life cycles stages (AXA 24 h onwards, inMAC 12 h onwards and FP 3 onwards, Fig 3B-3D respectively). Strong defects in both axenic amastigote differentiation and macrophage infection were observed for Δdub4, Δdub7 and Δdub13 (�85-fold, �7-fold and �50-fold decreases in fitness respectively). Additionally, decreased fitness was observed for Δdub5, Δdub8, Δdub9, Δdub10 and Δdub11 between the AXA 24 h and AXA 120h stages and for Δdub6, Δdub11 and Δdub17 between the META and inMAC 12 h stages. Consistent with their phenotypes in axenic amastigotes, Δdub4, Δdub7 and Δdub13 showed severe reductions (to zero) in fitness during mouse infection. Additional defects of 20-fold or more reduction in fitness were also seen for Δdub3, Δdub5, Δdub6, Δdub8, Δdub10, Δdub11 and Δdub14. While many examples of loss of fitness in mouse infection were observed for DUBs and other ubiquitination system mutants studied (S4 Table), loss of fitness phenotypes were less common for other peptidase mutants in the screen,
To demonstrate the validity of the phenotypes observed in the bar-seq screen, null mutants were also analysed individually at 24, 72 and 144 h after induction of amastigote differentiation and 48 h after transformation of axenic amastigotes back to promastigotes using resazurin, an oxidation-reduction indicator that reports on cell activity. The data from this experiment match well with the phenotypes observed in the AXA bar-seq data ( Fig 3E). An exception to this is Δdub9 which is well represented throughout the bar-seq experiment but loses viability at 72 h in the cellular assay, perhaps indicating the persistence of Δdub9 beyond 72 h but in a metabolically inactive state. The viability data also support more unusual phenotypes. For example, Δdub6 is still well represented at the AXA 24 h and 120 h stages of the bar-seq experiment but less so at the inMAC 12 h stage. This is supported by data from the viability assay showing that Δdub6 is viable throughout amastigote differentiation. Based on this finding, one could hypothesise that DUB6 may contribute to part of the infection process such as parasite entry into macrophages or survival in the phagolysosomes. Overall, the phenotypes observed in the pooled context of the bar-seq experiment reflect well the individual phenotyping of the viability assay. To further validate how these null mutants behave during amastigote differentiation, a western blot against HASPB, a marker for amastigotes [27], was performed. Δdub4, Δdub7, Δdub9 and Δdub13 showed very low or no expression of HASPB 120 h after the initiation of differentiation (S3 Fig). These results further demonstrate the inability of Δdub4, Δdub7, Δdub9, and Δdub13 to differentiate into amastigotes (first shown in Fig 3E).

DUB2 activity is essential for L. mexicana
L. mexicana DUB2 is one of the four essential DUBs identified in the CRISPR-Cas9 screen and is classified as a member of the USP family, closely related to mammalian USP5, USP13 and USP1 ( Fig 1B). DUB2 domain analysis suggested the presence of a peptidase C19, ubiquitin carboxyl-terminal hydrolase domain, a UBA (ubiquitin-associated domain and a Zinc finger UBP-type) (Fig 1A). The UBA domain found in the C-terminal end of DUB2 is present in many proteins associated with the ubiquitin system, and it is responsible for the binding of ubiquitinated substrates [28]. The Zinc finger domain is found only in a minor subfamily of USP DUBs including the mammalian USP5, USP13 and USP39, regulators of deubiquitination [18].
A modified version of the inducible DiCre system [29] was applied to L. mexicana DUB2 to investigate its function in promastigotes. The first allele of DUB2 was replaced with a DUB2 flanked by LoxP sites (Δdub2::DUB2 FLOX /DUB2, called DUB2 +/+FLOX ). Subsequently, the second allele of DUB2 was replaced with a HYG resistance cassette (Δdub2::HYG/Δdub2::DUB2- Excision of DUB2 in DUB2 +/+FLOX and DUB2 -/+FLOX logarithmic phase promastigotes was confirmed by the PCR amplification of a 1.3 kb DNA fragment in only those cells treated with 100 nM rapamycin ( Fig 4A). A pronounced defect in cell growth was observed in the rapamycin treated DUB2 -/+FLOX cell line in comparison to DUB2 +/+FLOX or untreated controls, indicating that deletion of both DUB2 alleles is required to elicit a growth phenotype (Fig 4B). The Cy5UbPRG ABP demonstrated that the activity of DUB2 sharply decreased in the induced DUB2 -/+FLOX line, but not in DUB2 +/+FLOX or WT (Fig 4C). For further analysis, the in-gel fluorescence band intensity was quantified for the DUB2 -/+FLOX line, showing one band (band 9) had a 70% depletion after rapamycin induction ( Fig 4D). Band 9 is the same size as that observed in the episomal over-expression of DUB2 ( Fig 2E) and correlates with the predicted mass of DUB2 protein covalently bound to the ABP (~100 kDa). A second unidentified DUB, labelled Unk 4, had a slight increase in activity after depletion of DUB2 in the DUB2 -/+flox line (Fig 4C and 4D). Loss of DUB2 may cause a compensatory change in activity for this DUB.
To confirm that the depletion of DUB2 leads to cell death a clonogenic assay was performed on DUB2 +/+FLOX and DUB2 -/+FLOX promastigotes following treatment with rapamycin or DMSO for 48 h. A >90% decrease in numbers of clones surviving for both DUB2 -/+FLOX clones was observed for cells treated with rapamycin, whilst there was no difference in the number of clones surviving in the DUB2 +/+FLOX line, despite the presence of rapamycin ( Fig 4E). These data confirm that the loss of DUB2 leads to death of promastigotes and that one allele of DUB2 can sustain growth. Finally, diagnostic PCR confirmed the presence of DUB2 in DUB2 -/+FLOX cells that survived rapamycin treatment, showing that if DUB2 is deleted, the cells were nonviable (S5B Fig). In support of these observations a significant increase in cell death was observed using propidium iodide staining after induction (S5C Fig).
Alignment of the C19 domain of DUB2 with other DUBs allowed the identification of C312 as the likely active site Cys. A point mutation was introduced into DUB2 in a pNUS plasmid to change the encoded Cys 312 to alanine, which would result in abolishment of DUB2 activity. pNUS-DUB2 and pNUS-DUB2 C312A plasmids were transfected into DUB2 -/+FLOX and an ABP assay with Cy5UbPRG used to confirm an increase of DUB2 activity in DUB2 -/+FLOX [DUB2] but not DUB2 -/+FLOX [DUB2 C312A ] ( Fig 4F). Over-expression of DUB2 was confirmed in both DUB2 -/+FLOX [DUB2] and DUB2 -/+FLOX [DUB2 C312A ] cell lines by western blot using an antibody raised against T. brucei DUB2; this cross-reacts with L. mexicana DUB2 ( Fig 4G). These data confirm that C312 is the active site Cys and is required for DUB2 activity. The loss of the DUB2 following rapamycin treatment resulted in a 90% decrease in the number of DUB2 -/+FLOX [DUB2 C312A ] clones in comparison to DUB2 -/+FLOX [DUB2] ( Fig 4E); indeed, expression of DUB2 in the DUB2 -/+FLOX line resulted in full complementation of the cell death phenotype. Taken together, these experiments strongly support that the deubiquitinase activity of DUB2 is required for the survival of L. mexicana promastigotes.
The DUB2 -/+FLOX line was also used to examine if DUB2 is essential for infection and survival in BALB/c mice. Mid-log phase WT and DUB2 -/+FLOX were grown in the presence or absence of rapamycin for 72 h and then inoculated into BALB/c mice. Efficient excision of DUB2 was confirmed by PCR ( Fig 5A). The mice infected with the rapamycin-treated WT or non-treated DUB2 -/+FLOX showed a steady growth in footpad size whereas rapamycin-treated mice DUB2 -/+FLOX did not develop a footpad lesion (Fig 5B), had 100-fold less parasites than non-treated in the footpad and could not be detected in the popliteal draining lymph nodes ( Fig 5C). The presence of a small number of parasites in the footpad of treated mice might be due to the presence of a subpopulation of cells that retained the floxed DUB2. This was supported by the finding that DUB2 could be detected by PCR in cells isolated from treated

PLOS PATHOGENS
Leishmania deubiquitinases DUB2 -/+FLOX lesions ( Fig 5D). In summary, DUB2 is required either for the establishment and/or maintenance of an infection in mice.

The DUB2 interactome
To gain more insight into the function of DUB2, the DUB2 interactome was investigated. The expression of active N-terminally tagged 3 x myc mNeo DUB2 was confirmed with the ABP and Western blotting (S6 Fig). DUB2 and an N-terminally tagged 3 x myc mNeo RDK2 control were immuno-precipitated from L. mexicana promastigote cell lysates and proteins identified by mass spectrometry. Parasites were cross-linked with dithiobis(succinimidyl propionate) to enable transient interactors to be captured, including possible substrates of DUB2. In the DUB2 samples, a total of 1269 protein groups were identified across 3 independent replicates, of these 818 were commonly identified in all three DUB2 replicates. Data were processed with an interaction scoring algorithm SAINTq to provide a score on the probability of a true interaction with DUB2. This analysis produced a list of 110 high confidence DUB2 interactors at <1% FDR (Fig 6, S5 Table). The list includes a number of proteins involved in ubiquitination, including polyubiquitin, three putative E2 ubiquitin-conjugating enzymes (LmxM.34.1300, LmxM.04.0680, LmxM.13.1580) and an ubiquitin-like modifier peptidase (LmxM.32.2260). Among the most enriched proteins was CYC12 (LmxM.36.5640), an L-type cyclin involved in spliced leader trans-splicing of pre-mRNA [30]. Other proteins involved in mRNA splicing were also enriched including SmD2, U5 snRNP and snRNP-B. Three members of the cleavage and polyadenylation specificity factor complex (CPSF) and a cleavage stimulation factor co-purified, suggesting a role for DUB2 in the processing of mRNA. Proteins involved in mRNA capping, CBP110 and CGM1, were also strongly enriched with DUB2. In addition, 19 of the interactors are either annotated as RNA binding or known to have a role in RNA metabolism. A smaller number of DUB2 interactors are implicated in transcription, however, this group includes the top enriched interactor NIF3. Interestingly, among this group of proteins involved in transcription/chromatin dynamics is Nucleosome Assembly Protein, NAP, which was enriched 27-fold. In humans, NAP1 is a histone chaperone and plays an important role in nucleosome assembly/disassembly by binding to free H2A/H2B dimers PCR amplification of DUB2 +/+FLOX and DUB2 -/+FLOX +/-100 nM RAP for 48 hours was conducted and the resulting amplicons resolved on 1% agarose gel stain with SYBR safe. (B) DUB2 +/+FLOX and DUB2 -/+FLOX promastigotes were seeded at a density of 1 × 10 5 cells mL −1 and grown in the absence or presence (-/+) of 100 nM RAP for 48 h. Cells were diluted to 1 × 10 5 cells mL −1 and allowed to grow in the presence or absence of 100 nM RAP for 4 d. Cell density was determined by counting at 24 h intervals and mean ± SD of triplicate values were plotted. The data were analysed in Prism software, and an unpaired t-test performed to indicate the significance (n = 5) between different time points, comparing treated with RAP to untreated samples. Error bars show the standard deviation of the mean. ���� p < 0.0001. (C) DUBs visualised by in-gel fluorescence analysis after separation by SDS PAGE. Samples were lysed, and 25 μg of lysate was incubated with the probe for 30 min. The positions of the different DUBs are indicated. (D) Fold difference of the DUB activity comparing DUB2 induced/DUB2 non-induced of the DUB2 -/+FLOX , data from (C). Band intensities were quantified by densitometry using GelAnalyzer 19.1 software. Each band is represented as a percentage of the total activity. The percentages for each band for the induced sample were divided by the non-induced to determine the fold difference. The data were then analysed in Prism software. Each column represents the mean of data from three independent experiments. Error bars show the standard deviation of the mean. Blue represents an increase of more than 1.3-fold whereas red represents a difference of less than 0.

PLOS PATHOGENS
Leishmania deubiquitinases [31]. Cycles of ubiquitination and de-ubiquitination of H2B are well known to play an important role in nucleosome dynamics during transcription and DNA replication. Several proteins with key roles in DNA replication were also identified as DUB2 interactors, such as PCNA, DNA polymerase delta and subunits of replication factor A and C. In other organisms, multiple components of the DNA replication machinery are ubiquitinated, in particular, ubiquitination of PCNA is critical for recruitment of DNA polymerases [32]. The identification of VPS4 [33] and a dynamin-1 like protein suggest involvement of DUB2 in endosomal trafficking.

Leishmania deubiquitinases
Finally, 3 protein kinases were co-enriched, including a mitogen activated protein kinase kinase (MKK1). This suggests a mechanism of DUB2 regulation, therefore mass spectrometry data were searched for phosphorylation sites revealing that DUB2 is phosphorylated on a serine in the region 586-589. A ubiquitination site in DUB2, K681, was also identified suggesting that DUB2 itself may be regulated by the ubiquitination pathway (S6 Table). Overall, the multiple interactions of DUB2 with splicing factors, proteins involved in mRNA capping and factors regulating chromatin dynamics/transcription, place DUB2 near actively transcribing genes.

DUB2 has deubiquitinating activity
To explore the function of DUB2 in more detail, recombinant enzyme (DUB2) and a catalytically inactive form (DUB2 C312A ) were expressed and purified using a baculovirus insect cell expression system (S7 Fig). DUB2, but not DUB2 C312A recombinant protein reacted with the Cy5UbPRG probe ( Fig 7A) and 85% of the recombinant DUB2 reacted with the probe ( Fig  7B). DUB2 linkage specificity against all eight types of human di-ubiquitin was then investigated over 30 min at a constant enzyme and substrate concentration. At 0.02 nM DUB2 had a preference for cleavage at Lys-11, Lys-29, Lys-33 and Lys-63, but at 2 nM DUB2 cleaved all linkages except Lys-27 ( Fig 7C). Finally, DUB2, but not DUB2 C312A recombinant protein had activity towards Ub-AMC substrate with a k cat of 0.049 s -1 and a K m of 14.9 x 10 −6 M (Fig 7D).

Discussion
Recent studies showed the potential for proteasome inhibition as a treatment for three diseases caused by trypanosomatid parasitic protozoa, including leishmaniasis [9,10]. The proteasome is dependent on the ubiquitination pathway to identify substrates for degradation and a search of the Leishmania genome reveals a range of E1, E2 and E3 ligases, as well as a number of DUBs [34]. Nevertheless, there have been no extensive studies exploring the role and essentiality of Leishmania DUBs through the complex life cycle of the parasite. Our bioinformatic analysis identified 20 DUBs belonging to the C12, C19 and C65 peptidase families, suggesting an extensive deubiquitinating system in L. mexicana, comparable in size to Saccharomyces cerevisiae [35]. Additional cysteine peptidase DUBs from C78, C85 and clan CP C97 families have been characterised in another study [26]. Analysis of DUBs in promastigote lysates using the Cy5UbPRG ABP probe identified 6 active DUBs (DUB2, DUB15, DUB16, DUB17, DUB18 and DUB19) by mass spectrometry and additional active enzymes by in-gel fluorescence, demonstrating the presence of an active deubiquitinating system in L. mexicana. Additional active DUBs were identified by in-gel fluorescence comparison of wild type and gene deletion mutants (DUB6, DUB9), yet many remain unidentified and could be active below the limit of

PLOS PATHOGENS
Leishmania deubiquitinases detection or possibly expressed in life cycle stages in the sandfly that were not examined in this study. We also used the ABP to demonstrate that many of the DUBs have activity that appears to remain constant in the different life cycle stages, yet we also demonstrated that Δdub4, Δdub7 and Δdub13 had no detectable loss of fitness in the promastigote, but had a significant loss of fitness in the amastigote. This apparent contradiction might be explained by a specific requirement for stage-regulated DUB substrates. Intracellular amastigotes have a stringent metabolic response that is acutely sensitive to metabolic perturbation [36]. Our data indicate that amastigotes are also highly sensitive to disruption of ubiquitination homeostasis.
The bar-seq screen and individual viability assay of DUB null mutants highlighted DUBs with potential roles in differentiation and infection. Δdub4, Δdub7 and Δdub13 had the strongest loss of fitness defects observed in axenic amastigote, macrophage infection and mouse infection experiments. Additionally, Δdub3, Δdub5, Δdub6, Δdub8, Δdub10, Δdub11 and Δdub14 showed strong loss of fitness uniquely during mouse infection, perhaps due to increased selective pressures inside the host or the longer duration of the experiment. Although we identified a number of life cycle defects using the bar-seq screen, there are some potential limitations to using a pooled library, as the fitness of individual null mutants may be influenced by other mutants in the pool. This may explain, for example, the different phenotypes observed for Δdub9 between the bar-seq experiment and viability assay/HASPB blot. Our individual profiling of axenic amastigote null mutants using the viability assay confirms the phenotypes we observed in the bar-seq screen, but further individual assessment on selected null mutant lines could be advantageous. The identification of several DUBs essential for differentiation was not unexpected as the Leishmania proteome differs substantially between promastigote and amastigote stages [37]. DUBs can control the abundance of Leishmania proteins by rescuing them from proteasomal degradation or, more indirectly, by regulating autophagy. Autophagy has been shown to be essential for the differentiation of promastigotes into amastigotes both in vitro and in vivo [33,38]. Proteolysis is required for metacyclogenesis and the promastigote to amastigote transition [39] and an accumulation of ubiquitylated proteins in the cytosol has been documented at the onset of differentiation [40]. Therefore, a role for Leishmania DUBs in differentiation-associated protein degradation could explain the higher degree of essentiality observed for DUBs over other peptidases in our barseq screen.
We were unable to generate null mutant lines for four DUBs, three C19 (DUB1, DUB2 and DUB12) and one C12 family (DUB16) peptidase. The successful generation of facilitated null mutant lines for the four DUBs provides a level of support that they are essential for the promastigote. DUB16 is related to human UCHL1 (33% identity) and UCHL3 (37% identity), for which one proposed role is the control of free ubiquitin levels through the processing of ubiquitin precursor and ubiquitinated proteins [41]. DUB16 could have a role in regulating free ubiquitin levels, something that must be further investigated.
USP36 is the human DUB most closely related to DUB12 (28% identity) and has been shown to regulate the stability of nuclear proteins including c-Myc, B23 and fibrillarin, the latter two of which are required for rRNA processing and ribosomal biogenesis [42]. USP36 deletion is lethal in mice, attributable to its role in rRNA processing and protein synthesis [43]. However, deletion of the yeast homologue, Ubp10, only results in a reduced growth phenotype [44]. Further investigation must be performed to investigate if DUB12 shares any functional similarity with USP36 and to explain how DUB12 depletion caused cell death. Lastly, for the DUB12 overexpression line, we observed an increase in the activity of two forms of DUB12, likely arising from its proteolytic cleavage to 55 kDa proteins from an 85 kDa precursor. The post-translational modification of DUBs is one way in which their activity has been shown to be regulated and, although rare, proteolytic cleavage of DUBs has been described for some human DUBs [45,46]. DUB2 is a cysteine peptidase with a C19-UCH domain, a UBA domain at the C-terminus and a ZnF_UBP domain at the N-terminus. The most similar human DUBs are USP5 and USP13. USP5 is a well characterised member of the USP family responsible for the disassembly of the majority of unanchored poly-ubiquitin in vivo [47]. The ZnF-UBP domain of USP5 recognises the C-terminus of ubiquitin [48] and USP5 has been shown to hydrolyse Lys-63, Lys-48, Lys-11, Lys-29 linkages [49]. DUB2 is similar to USP5 in having broad linkage specificity towards all the di-ubiquitin chains except Met-1 and Lys-29 in vitro. Our in vitro experiments demonstrate that DUB2 exhibits deubiquitinating activity against both artificial DUB substrates (Cy5-Ub-PRG ABP and Ub-AMC) and di-ubiquitin. Several human DUBs have previously been demonstrated to have a pleiotropic function by regulating protein stability or by regulating assembly and function of different machineries. Similarly, the essentiality, localisation and interactome of DUB2 suggests a pleiotropic role in the deubiquitination of multiple substrates. The possibility that DUB2 is responsible for the release of ubiquitin from polyubiquitin chains could also explain the phenotype in Leishmania after DUB2 depletion. The sudden and significant depletion of free ubiquitin results in significant dysfunction of all molecular machinery associated with ubiquitin, such as proteasomal degradation, endocytosis, replication, transcription and RNA splicing. If this was the case, the parasite would enter a 'free ubiquitin crisis', which would eventually lead to death.
The interactome data suggest DUB2 plays a role in post-transcriptional gene regulation. In Leishmania and other kinetoplastids, gene expression occurs in an unusual manner involving constitutive polycistronic transcription and trans-splicing of a splice leader RNA to generate mature mRNAs [50]. Amongst the DUB2 interactors, one of the most strongly enriched proteins was cyclin 12 which in trypanosomes forms a tripartite complex with CRK9 and a CRK9-associated protein, CRK9AP. The CYC12-CRK9-CRK9AP complex is essential for trans-splicing of the splice leader RNA [30]. However, CRK9 and CRK9AP were not detected in the DUB2 interactome, indicating that DUB2 may play a role in turnover of free CYC12. Depletion of CRK9AP causes a rapid loss of CYC12 at the protein but not mRNA level, suggesting dynamic regulation of CYC12 protein [30]. DUB2 may therefore protect free CYC12 from degradation. When DUB2 is absent, CYC12 may become a target for degradation, leading to co-depletion of CRK9 and a critical loss of trans splicing capacity. Alternatively, deubiquitination of CYC12 may be required for complex formation, similar to the requirement for the removal of ubiquitin from Bcl10 by Usp9x for its tripartite complexing to Malt1 and Carma1 [51]. Another strongly enriched protein in the DUB2 interactome was the histone chaperone NAP1, an important factor in nucleosome assembly [31]. This implies a role for DUB2 in nucleosome dynamics. In other eukaryotes, ubiquitination/de-ubiquitination of H2A and H2B is a dynamic process that modulates the stability of nucleosomes to allow transcription/DNA replication to take place [52]. The transcription factors, splicing factors and DNA replication proteins co-enriched with DUB2 are consistent with a role in regulating nucleosome assembly/disassembly since this would place DUB2 in the vicinity of these processes.
DUB activity can be regulated by post-translational modifications such as phosphorylation, for which one of the best characterised examples is activation of OTUD5/DUBA (OTUdomain-containing protein 5) by phosphorylation in the catalytic domain [53]. Ubiquitination can also affect DUB activity, for example, ubiquitination of UCH-L1 near the active site prevents binding to ubiquitin [54]. We identified a phosphorylated region and a ubiquitination site flanking the DUB2 UBA domain which could similarly regulate binding of DUB2 to ubiquitinated substrates. DUB2 has a broad specificity for diubiquitin cleavage, with Lys-27 being the only linkage for which there is poor activity. Lys-11, Lys-29 and Lys-48 linked chains have roles in proteasomal degradation, Lys-29 is associated with lysosomal degradation and Lys-33 is involved in both secretory and endocytic pathways, acting as a sorting signal for membrane proteins [55]. Another preferred linkage of DUB2 is Lys-63 which can serve as a molecular glue, allowing the rapid and reversible formation of pivotal signalling complexes. Lys-63 ubiquitin chains have been shown to regulate several biological processes including DNA repair, clearance of damaged mitochondria and protein sorting and they can guide assembly of large protein complexes that drive mRNA splicing and translation [55]. In the DUB2 pull down, the VPS4 protein was identified, which could indicate a putative role of DUB2 in endosome sorting and could explain why DUB2 cleaves Lys-33 and Lys-29. Whilst VPS4 is the only known protein associated with autophagy/endocytosis that was identified in the DUB2 interactome, a role in the endocytic system fits with DUB2 lysosomal localisation. Previously, VPS4 was shown to play a role in endosome sorting and autophagy in L. major, which are essential for the differentiation and virulence of this parasite [33].
In the past decade there has been an expanding interest in DUBs as potential drug targets as they have been shown to play a crucial role in both cancer and neurodegenerative diseases [56]. The determination of the structure of multiple DUBs, including catalytic domains, in combination with the development of advanced high-throughput screening-compatible assays with USP substrates has resulted in the discovery of several inhibitors against a select number of human DUBs including USP7 [57,58] and USP14 [59]. DUB2 is the first DUB to have been genetically validated as a drug target in L. mexicana. DiCre permitted the analysis of emerging phenotypes, confirming the essential role for DUB2 in promastigotes and in the establishment of mouse infection. Analysing Leishmania genes in this manner represents a high level of genetic validation [60]. To date, most drug target validation has relied on genetic manipulation of promastigotes, however, the bar-seq approach presented here has revealed some DUBs to have amastigote-specific loss of fitness, making DUBs 3, 5, 6, 8, 10, 11 and 14 potential drug targets. As DUB6 is related to human USP7 and DUB14 to human USP14, chemical tools developed for human DUB drug discovery programs [59] may be useful chemical entry points to develop inhibitors to the Leishmania enzymes.

Ethics statement
All experiments were conducted according to the Animals (Scientific Procedures) Act of 1986, United Kingdom, and had approval from the University of York Animal Welfare and Ethical Review Body (AWERB) committee.
L. mexicana promastigotes were differentiated to axenic amastigotes as previously described [61]. Briefly, stationary phase culture of L. mexicana was pelleted by centrifugation at 1,000 x g for 10 minutes and washed with 1 x PBS. Cells were then resuspended in amastigote culture medium (Schneider's Drosophila medium [Gibco], 20% HIFCS and 0.015 mg mL -1 hemin [stock dissolved in 50 mM NaOH], pH 5.5) with a concentration of 2 x 10 6 cells mL -1 in 6-well plates. Except where stated otherwise, cells were then incubated at 32˚C with 5% CO 2 . For transformation of amastigotes back to promastigotes after 120 h, 100 μL of axenic amastigotes
To assess infection of mice, 10 6 stationary phase L. mexicana promastigotes were injected into the right footpad of female BALB/c mice (Charles River Laboratories). Lesion size was monitored weekly. Parasite burden was assessed as previously described [62]. For profiling of DUBs in lesion-derived amastigotes, 4 x 10 6 stationary phase L. mexicana were injected into the rump of female BALB/c mice. Lesion size was monitored weekly. For the purification of lesion derived amastigotes the lesion tissue was collected in amastigote stabilisation buffer (PSGEMKA) [63], cut into pieces and passed through a sterile 100 μm cell strainer (BD) to disrupt the host macrophages and liberate the amastigotes. The suspension was incubated with 125 mg L -1 saponin for 5 mins, centrifuged 2,000 x g for 10 mins and the pellet was washed twice with PSGEMKA buffer. Cell debris was eliminated by passage through a glass wool-Sephadex CM-25 column. The suspension obtained was centrifuged at 2,000 x g for 10 mins. Thereafter, the supernatants were discarded, and the pellets were resuspended in Schneider's Drosophila medium supplemented as indicated above.
L. mexicana transfections were performed using a T Cell Nucleofector Kit (Lonza) as described previously [26]. The transfected culture was split between two flasks to select for independent transfection events and incubated overnight at 25˚C to recover. The next day, suitable antibiotics were added and cells cloned by serial dilutions of 1 in 6, 1 in 66 and 1 in 726 in 96-well microplates.

Phylogenetic analysis
The protein sequences for DUBs were extracted from (Uniprot-https://www.uniprot.org/) [64]. For human DUBs the canonical sequence was used as it was determined by Uniprot. The protein alignment was performed using MUSCLE (multiple sequence alignment with high accuracy and high throughput). A constraint tree was created by including DUBs into the seven known families and a phylogenetic tree was generated by RAxML (https://raxml-ng. vital-it.ch/#/) [65]. The LG Substitution matrix was used. The best fit model tree was further designed initially in the iTOL INTERACTIVE TREE OF LIFE (https://itol.embl.de/) where branch length was ignored, and an unrooted tree style was formed. Finally, Adobe Illustrator was then used to finalise the tree.

Life cycle phenotyping using bar-seq
Null mutants were generated using a CRISPR-Cas9-based approach [23] with a cell line expressing Cas9 and T7 RNA polymerase [26]. DNA was harvested from clones or populations for diagnostic PCRs using the QIAGEN DNeasy Blood and Tissue Kit. Two sets of primers were designed to confirm the generation of null mutants, the first specific to the ORF of the target gene and the second specific to the UTRs (forward primer 5'-UTR, reverse primer 3'-UTR).
Null mutant lines were grown to mid-log phase, spun down at 1,000 x g for 10 min and pooled in equal proportions in Grace's medium (Grace's insect media [Sigma-Aldrich] with

PLOS PATHOGENS
Leishmania deubiquitinases 4.2 mM NaHCO 3 , 10% FCS [Gibco], 1 x Penicillin/Streptomycin [Gibco] and 1 x BME Vitamins [Sigma-Aldrich] and adjusted to pH 5.25) up to a total concentration of 2 x 10 6 cells mL -1 . Cultures were prepared in sextuplicate to provide technical replication. Next, the pooled cells were grown at 25˚C for 7 d to allow for the enrichment of metacyclic promastigotes. To prepare axenic amastigotes, promastigotes from the 168 h time point were centrifuged and resuspended in amastigote medium at 2 x 10 6 cells per well. Cells were then cultured at 35˚C with 5% CO 2 .
Bone marrow-derived macrophages were extracted from female BALB/C mouse femur, equilibrated in warm DMEM (GE Life Sciences), spun at 200 x g for 10 min and resuspended in macrophage medium (DMEM [Invitrogen] plus 10% FCS and 2 mM L-glutamine) in order to have 2.5 x 10 6 cells per well [66]. Enriched metacyclic cells were purified using a Ficoll 400 (Sigma-Aldrich) gradient (40%, 10% and 5% Ficoll) and centrifuged at 1,300 x g for 10 min. Following washing with Grace's medium, metacyclic promastigotes were added to the macrophages in a 1:1 ratio and left to interact for 4 h at 35˚C with 5% CO 2 . Excess Leishmania cells were then removed by washing with DMEM and the cells incubated in DMEM 10% HIFCS for 12 h and 72 h.
To perform the mouse infections, 10 6 purified metacyclic promastigotes were injected into the left footpad of BALB/c mice. Mice were culled at 3 and 6 weeks post-inoculation. At the relevant time points, DNA was extracted from experimental samples and processed using the QIAGEN DNeasy Blood and Tissue Kit. DNA samples were prepared for next-generation sequencing by first amplifying unique barcodes by PCR using the following primers: 5'-TCG TCGGCAGCGTCAGATGTGTATAAGAGACAGAgatgatgattacTAATACGACTCACTA TAAAACTGGAAG-3' and 5'-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGA GAGACAGGCATGCCTT-3' containing the Illumina adapter sequence. Reactions were set up with Q5 polymerase (NEB) as per the manufacturer's instructions except for the following changes: just 0.5 x of Q5 Reaction Buffer and Q5 High GC Enhancer were used. The cycling conditions were 98˚C for 5 min followed by 20-25 cycles of 98˚C for 30 sec, 60˚C for 30 sec and 72˚C for 10 sec. Following purification of the PCR reactions using the QIAGEN MinElute PCR Purification Kit, a second PCR was performed using Illumina's Nextera XT indexing primers in order to add unique barcode sequences to each sample. 8 cycles of PCR amplification were performed using NEBNext Q5 Polymerase 2X Master Mix (New England Biolabs), according to the manufacturer's guidelines. Indexed amplicons were purified using 0.9 X AMPure XP beads (Beckmann Coulter) and eluted into low TE buffer before quantitation and pooling at approximately equimolar ratios. Amplicon pools were then sent for 150 base paired end sequencing on an Illumina HiSeq 3000 sequencer at the University of Leeds Next Generation Sequencing Facility.
Illumina read sequences were analysed using a custom Python script to search for the 12 bp sequence preceding the barcode. If an exact match was found, the following 12 bp sequence was counted as a barcode sequence. A total count for each unique barcode sequence was generated for each sample. Next, the relative fitness of null mutants at each life cycle stage was calculated by taking the number of unique barcode reads for each null mutant line and dividing it by the total number of expected barcode reads. Significant increases or decreases in null mutant line fitness between adjacent samples in the experimental workflow were analysed using unpaired t-tests and the Holm-Šídák method in GraphPad Prism 7.

Construction of plasmids
A full list and descriptions of all primers and plasmids used in this study can be found in S3 Table. The construction of the plasmids needed for the generation of DUB2 inducible gene

PLOS PATHOGENS
Leishmania deubiquitinases deletion was performed as previously described [29]. The DUB2 gene including stop codon was cloned into pGL2315, creating pGL2727. For the complementation plasmids, pNUS-DUB1, DUB2, DUB12 and DUB16 expression plasmids were generated using Gibson Assembly. The successful clones were confirmed using PCR and sequencing.

Induction of diCre mediated gene deletion
A modified version of the DiCre inducible system [29] was used. Two dimerisable Cre recombinase subunits were integrated into the ribosomal locus. DiCre in cell lines containing both diCre and loxP flanked genes of interest was induced by the addition of between 100 nM and 1 μM rapamycin (Abcam). Promastigotes were then allowed to grow for 48 h and split into new cultures with a concentration of 1 x 10 5 cells mL -1 and induced again with 100 nM of rapamycin.

Viability analyses
2-5 x 10 6 L. mexicana promastigotes were pelleted by centrifugation at 1,000 x g for 10 mins and washed twice with PBS with 5mM of EDTA. Cells were resuspended in 1mL of PBS with 5mM EDTA and 1μg mL -1 of propidium iodide (PI) and 1 μg mL -1 of RNAse A. Cell fluorescence was analysed on a Beckman Coulter, CyAn ADP. Data were analysed on FlowJo software (Tree Star Inc.).

Preparation of protein extracts
3 x 10 7 L. mexicana cells in the desired life cycle stage were pelleted by centrifugation at 1,000 g for 10 min. Cells were washed in 1 x PBS and lysed by resuspending in 40 μL of 1 x SDS-PAGE loading buffer. Next, samples were heated for 10 minutes at 90˚C on a heat block. Samples were allowed to cool down at room temperature and 250 U of Expedeon BaseMuncher endonuclease was added to each sample. Samples were then incubated at 37˚C for 30 minutes and loaded directly into a NuPAGE 4-12% Bis-Tris protein gel (Thermofisher) or stored at -20˚C.
Cy5UbPRG profiling 3 x 10 7 cells from a log-phase L. mexicana culture or 9 x 10 7 amastigote were spun at 1,000 x g and washed three times with 1 mL of ice-cold TSB Wash buffer (44 mM NaCl, 5 mM KCl, 3 mM NaH 2 PO 4 , 118 mM sucrose and 10 mM glucose, pH 7.4). Next, the cells were lysed using a newly prepared ice-cold lysis buffer (50 mM Tris-HCl pH 7.4, 120 mM NaCl, 1% NP40 and freshly added in order: 1 μg mL -1 pepstatin, 1 x cOmplete ULTRA Tablets, [Mini, EASYpack Protease Inhibitor Cocktail, Roche], 1 mM DTT, 1 mM PMSF, 0.01 mM E64). Samples were incubated at 4˚C for 15 minutes. Afterwards, samples were centrifuged at 17,000 x g for 15 minutes and the supernatant withdrawn. Samples were prepared to have a protein concentration of 1 mg mL -1 in a total volume of 25 μL. 2 μL of 50 mM NaOH and 1 μM of Cy5UbPRG (UbiQ) were added. Lysis buffer was used to top up to a final volume of 25 μL. The reaction was then incubated at room temperature for 30 min and stopped by the addition of 3 x loading dye. 12 μL were then analysed in a gradient 4-12% NuPage Bis-Tris protein gel, for exactly 90 min at 200 V. The gel was then washed with water and imaged with Amersham Typhoon with Excitation: 635 nm, Filter: Cy5 670BP30 (GE Healthcare Life Sciences).
For purification of DUBs cyanogen bromide (CNBr)-activated Sepharose 4B resin (GE Healthcare) was used to couple with Ub-PRG (1.2 pmol) or Ubiquitin (1.2 pmol) according to the manufacturer's protocol (GE Healthcare). Beads were then stored in 4˚C until used. 50 mg of the coupled resin of either Ub-PRG or ubiquitin were incubated with fresh Leishmania lysate collected from 1 x10 8 cells (prepared as described above) at 37˚C for 3 h. The beads were then washed with PBS plus 1% Triton X-100 (3 x 10 mL), 2% CHAPS, 8 M Urea in PBS and distilled water. The bound proteins were then digested on-bead with trypsin and analysed by MS/MS spectrometry. The MS data was processed using Data Analysis software (Bruker) and the automated Matrix Science Mascot Daemon server (v2.4.1). Protein identifications were assigned using the Mascot search engine to interrogate protein sequences in the annotated proteins database (obtained from TriTrypDB), allowing a mass tolerance of 0.4 Da for both MS and MS/MS analyses.

Immunoprecipitation of myc-tagged Leishmania proteins
For each of a total of 3 independent replicates, 1 x 10 9 parasites were washed twice in PBS and cross-linked with 1 mM dithiobis(succinimidyl propionate) (DSP, Thermo Scientific) in 10 ml PBS for 10 min at 26˚C. DSP was quenched with 20 mM Tris pH 7.5 for 5 min. Parasites were lysed in ice cold 500 μl lysis buffer (1% NP40, 50 mM Tris pH 7.5, 250 mM NaCl, 1 mM EDTA, 0.1 mM PMSF, 1 μg mL -1 pepstatin A, 1 μM E64, 0.4 mM 1-10 phenanthroline, 20 μl mL -1 proteoloc (Expedeon), 0.17 complete protease inhibitor tablets mL -1 (Sigma)) by probe sonication for 3 × 5 sec. Lysate was centrifuged for 10 min at 10,000 x g at 4˚C and protein concentration in the supernatant measured by BCA assay (Thermo scientific). Lysate equivalent to 5 mg of total protein was incubated with 30 μl anti-myc magnetic beads (Thermo scientific) for 2.5 hr at 4˚C with rotation. Beads were washed 4 × 300 μl ice cold lysis buffer for 5 min each wash followed by two PBS washes. Beads were then stored at -80˚C. Myc-tagged protein was eluted at room temperature with 25 μl myc-peptide (0.5 mgmL -1 in PBS, Sigma) for 15 minutes at 700 rpm mixing. The elution step was repeated and eluates pooled. 1/10 th of the elution was used to check successful immunoprecipitation by western blot. The remainder was mixed with 4 volumes of absolute methanol and 1 volume of chloroform and vortexed for 1 min. The sample was then centrifuged for 1 h at 18,000 x g at 4˚C. The pellet was washed with 270 μL of absolute methanol and centrifuged for 10 min at 18,000 x g at room temperature. The pellet was resuspended in 150 μL 50 mM TEAB pH 8.5, 0.1% PPS silent surfactant (Expedeon) for 1 h by shaking at 800 rpm at room temperature. Afterwards, 10 mM of Tris (2-carboxyethyl) phosphine (TCEP) and 10 mM Iodoacetamide (IAA) were added to the sample and incubated for 30 min at room temperature in the dark. Finally, 200 ng of trypsin and 1 mM CaCl 2 were added and proteins digested overnight at 37˚C 200 rpm. After digestion, PPS silent surfactant was cleaved by acidifying the digest to 0.5% trifluoroacetic acid (TFA) and incubating 1 h at RT. Digest was centrifuged for 10 min at 17,000 x g. Peptides were desalted with C18 (3M Empore) desalting tips.
Samples were loaded onto an UltiMate 3000 RSLCnano HPLC system (Thermo) equipped with a PepMap 100 Å C18, 5 μm trap column (300 μm x 5 mm Thermo) and a PepMap, 2 μm, 100 Å, C18 EasyNano nanocapillary column (75 μm x 150 mm, Thermo). The trap wash solvent was aqueous 0.05% (v:v) trifluoroacetic acid and the trapping flow rate was 15 μL min -1 . The trap was washed for 3 min before switching flow to the capillary column. Separation used gradient elution of two solvents: solvent A, aqueous 1% (v:v) formic acid; solvent B, aqueous 80% (v: v) acetonitrile containing 1% (v:v) formic acid. The flow rate for the capillary column was 300 nL min -1 and the column temperature was 40˚C. The linear multi-step gradient profile was: 3-10% B over 7 min, 10-35% B over 30 min, 35-99% B over 5 min and then proceeding to wash with 99% solvent B for 4 min. The column was returned to initial conditions and re-equilibrated for 15 min before subsequent injections. Each independent replicate was injected once.
The nanoLC system was interfaced with an Orbitrap Fusion Tribrid mass spectrometer (Thermo) with an EasyNano ionisation source (Thermo). Positive ESI-MS and MS2 spectra

PLOS PATHOGENS
Leishmania deubiquitinases were acquired using Xcalibur software (version 4.0, Thermo). Instrument source settings were: ion spray voltage, 1,900 V; sweep gas, 0 Arb; ion transfer tube temperature, 275˚C. MS1 spectra were acquired in the Orbitrap with: 120,000 resolution, scan range: m/z 375-1,500; AGC target, 4e5; max fill time, 100 ms. Data dependent acquisition was performed in top speed mode using a 1 sec cycle, selecting the most intense precursors with charge states >1. Easy-IC was used for internal calibration. Dynamic exclusion was performed for 50 sec post precursor selection and a minimum threshold for fragmentation was set at 5e3. MS2 spectra were acquired in the linear ion trap with: scan rate, turbo; quadrupole isolation, 1.6 m/z; activation type, HCD; activation energy, 32%; AGC target, 5e3; first mass, 110 m/z; max fill time, 100 msec. Acquisitions were arranged by Xcalibur to inject ions for all available parallelisable time.
Peak lists in.raw format were imported into Progenesis QI (Version 2.2., Waters) and LC-MS chromatograms aligned. Precursor ion intensities were normalised against total intensity for each acquisition. A combined peak list was exported in.mgf format for database searching against the L. mexicana subset of the TriTrypDB database (8,250 sequences; 5,180,224 residues). Mascot Daemon (version 2.6.1, Matrix Science) was used to submit the search to a locally-running copy of the Mascot program (Matrix Science Ltd., version 2.6.1). Search criteria specified: Enzyme, trypsin; Max missed cleavages, 2; Fixed modifications, Carbamidomethyl (C); Variable modifications, Oxidation (M), Phosphorylation (S,T,Y); Peptide tolerance, 3 ppm; MS/MS tolerance, 0.5 Da; Instrument, ESI-TRAP. Peptide identifications were passed through the percolator algorithm to achieve a 1% false discovery rate as assessed against a reversed database and individual matches further filtered to minimum expect scores of 0.05. The Mascot.XML result file was imported into Progenesis QI and peptide identifications associated with precursor peak areas. Relative protein abundance was calculated using precursor ion areas from non-conflicting unique peptides. Accepted protein quantifications were required to contain a minimum of two unique peptide matches. Interactors were scored using SAINTq [67], using a false discovery rate threshold of <1% to select high confidence DUB2 interactors. Complete mass spectrometry data sets are available to download from Mas-sIVE (MSV000085242) and ProteomeXchange (PXD018415). The doi for the data is [doi:10. 25345/C5Z10J].

Protein expression and purification
A baculovirus expression system was used for the expression of DUB2 recombinant protein in sf9 insect cells as described in the Bac-to-Bac Baculovirus Expression protocol (Invitrogen). The gene was first cloned into the PFastBacNKI-his-3C-LIC vector using Ligation Independent Cloning (LIC). After three days, the insect cells were lysed in lysis buffer (30 mM Tris-HCl, pH 8, 0.3 M NaCl, 0.03 M imidazole, 5 mM β-mercaptoethanol and 1 tablet of cOmplete ULTRA Tablets, Mini, EASYpack Protease Inhibitor Cocktail, Roche) by sonication. The samples were then centrifuged (21 K, 4˚C, 30 min) and the supernatant applied to an AKTA Start (GE Life Sciences) at 5 mL min -1 , to a 5 mL His-trap crude FF column (GE Life Sciences). This was washed with Buffer A (30 mM Tris-HCl pH 8, 0.3 M NaCl, 0.03 M imidazole, 5 mM β-mercaptoethanol) and the protein eluted with increasing concentration of imidazole (0.03-0.5 M). Protein was concentrated using Amicon Ultra-4 30k MWCO (Millipore). 3C protease was added to purified protein in a 1:50 (w:w, 3C:Recombinant protein) ratio with 2 mM DTT. The aliquot was added to a 16 kDa dialysis tube and incubated at 4˚C in buffer A overnight. Protein was loaded onto a 5 mL His-trap crude FF column and ran as described above. The flow-through from the His-tag cleavage containing the tag free protein was concentrated to 2 mL using Amicon Ultra-4 30k MWCO (Millipore). The concentrated protein was applied to a HiLoad 16/600 S75pg column (GE Life Sciences), equilibrated and eluted in 25 mM Tris-HCl pH 8.0, 150 mM NaCl, 5 mM β-mercaptoethanol. Fractions containing the protein were identified by absorbance at 280 nm and analysed by 10% SDS-PAGE. These fractions were pooled and concentrated up to 2 mg mL -1 . 1 mM of DTT was added to the sample and the protein stored at -80˚C.
Ubiquitin probe assay for recombinant protein 10 μL of the appropriate concentration of the recombinant DUB was incubated with 2 μL 50 mM NaOH, 5 μL dH 2 O and 2 μL 0.25 μg mL -1 ubiquitin probe Cy5UbPRG (UbiQ) for 30 min at room temperature. Binding of probe was analysed using a NUPAGE 4-12% Bis-Tris Gel. The gel was then scanned using the Amersham Typhoon (GE Healthcare Life Sciences) and stained with InstantBlue Protein Stain.
The in vitro enzymatic assay with the fluorogenic substrate ubiquitin-7-amido-4-methylcoumarin (Ub-AMC, UbiQ) was assembled in a 384-well microplate (Thermo Scientific). Assays were performed in 20 μL reaction volumes and in triplicate. The reactions were initiated by the addition of 10 μL of substrate or enzyme. Substrate and enzymes were diluted in buffer containing 50 nM Tris, pH 8.0, 0.15 M NaCl, 5 nM DTT and 0.05% CHAPS. The fluorescence intensity was monitored with a Polarstar plate reader (Omega) equilibrated at 25˚C at intervals of 20 sec for 20 min using excitation/emission filter pairs of 340/460 nm. The fluorescence value was then converted into concentration using a standard curve. The standard curve was determined by allowing the reactions to plateau and the maximum arbitrary fluorescence units were measured and plotted against a known substrate concentration. An AMC concentration against time graph was generated in Excel software and the initial slope of each reaction was determined by generating a linear regression. For the determination of the Km and kcat values of DUB2, the enzyme concentrations were kept constant, and the concentration of Ub-AMC varied from 0.25 μM to 32 μM. The initial rates in μM sec -1 were then plotted against substrate concentration (μM) in Prism 7 software. The data were fitted to the Michaelis-Menton equation so the kinetic parameters, Km and kcat were determined. Diagnostic PCRs were performed to check the successful generation of null and facilitated null mutants. PCR analysis of genomic DNA of at least two individual clones (null mutants, CL1, CL2 etc) or two separate populations (facilitated null mutants), using the indicated set of primers: ORF-specific primers (A), UTRspecific primers (B) or ORF-specific, 5'-UTR-blasticidin or UTR-specific primers (C). As a control, the parental Cas9 T7 cell line was used. The resulting amplicons were resolved on a 1% agarose gel and stained with SYBR safe. The expected size of the resulting amplicons is presented in S3 Table. A schematic representation of the wild type, heterozygous, null mutant and facilitated null mutant genomic loci including the diagnostic primers (green arrow primers bind to ORF, grey arrow primers bind to UTRs and amplify across the loci, purple arrow

PLOS PATHOGENS
Leishmania deubiquitinases primers binds to the blasticidin resistance marker) is also shown in (C). (TIF)

S3 Fig. Activity profiling of the DUB null mutant cell lines. (A)
Lysate extracted from logphase L. mexicana promastigotes treated with or without Cy5UbPRG for 30 min. Proteins were separated by SDS-PAGE and in-gel fluorescence (Cy5) was captured using a Typhoon imager followed by Coomassie staining as a loading control. (B) Lysates extracted from null mutant lines of log-phase L. mexicana promastigotes treated with Cy5UbPRG for 30 min. Ingel fluorescence images were obtained as for (A). The red arrowhead shows the position where an active DUB is missing compared to the parental Cas9 T7 cell line. (C) Western blot analysis of 2 x 10 7 axenic amastigotes. Samples were separated in a 4-15% protein gel. The stain-free gel used contains trihalo compounds which, in the presence of the UV-light, react with tryptophan residues, producing fluorescence. The gel was activated by 45 sec UV exposure, proteins were transferred to a PVDF membrane and probed with 1:1,500 dilution of anti-HASPB. Finally, as a loading control the total protein was determined using the stain-free property of the gel. (D) Lysate extracted from differentiated promastigotes to axenic amastigotes (48 h and 144 h after initiation of axenic differentiation) treated with or without Cy5UbPRG for 30 min. Protein was separated in an SDS-PAGE gel, and the image was captured using a Typhoon imager and the gel stained with Coomassie. and DUB2 -/+FLOX [DUB2 C312A ] surviving parasites after treatment with RAP in the clonal assay ( Fig 4E). Schematic representation of the primers used for the PCR as well as the expected fragments are shown in Fig 4A. (C) Compiled flow cytometry results of live/dead cells detected using propidium iodide staining in DUB2 +/+flox and DUB2 -/+flox promastigotes over the course of 96 hours post induction. Cells were grown for 48 hours with or without rapamycin. Afterwards, the cells were seeded at a density of 1 × 10 5 cells mL −1 and allowed to grow in the presence or absence of 100 nM of RAP. Samples were collected every 24 h. Flow cytometry was used to analyse the samples. The data were then collected and analysed with FlowJo_v10 software, where the percentage of dead cells was determined. The data were analysed in Prism software and an unpaired t-test performed to indicate significance (n = 3) between different time points, comparing treated with RAP to untreated samples. Error bars indicate the standard deviation of a mean. ��� p < 0.001, ���� p < 0.0001.